Do announcements by credit rating agencies affect the sentiment of individual investors?
erasmus university rotterdam
Faculty:Behavioral Economics
Master:Financial Economics
Author: Ewoud van Leeuwen
Supervisor: Saskia ter Ellen
Student nr: 305092
Key words: investor sentiment, credit rating agencies, event study and financial crisis.
Abstract:
This paper investigates a relationship between credit rating agencies and investor sentiment. This paper researches whether investor sentiment changes upon an announcement by a credit rating agency on the sovereign credit worthiness. We make use of a dataset that was provided by one of the major Dutch banks, which gives us insight on the sentiment of preferred investors. Maximum – minimum spread, skewness and cross sectional dispersion in the expectations on the AEX are used as proxies for investor sentiment. The data of the three major credit rating agencies are used: Fitch, Moody’s and Standard&Poors. The results indicate that announcements by credit rating agencies on the sovereign credit rating affect the average sentiment of investors participating in our survey group.
1. Introduction
2. Theoretical evidence and institutional background
3. Methodology
Event study methodology
Testing Maximum – Minimum spread
Testing cross sectional dispersion
Testing kurtosis and skewness
Collecting credit rating announcement data
Testing investor sentiment proxy by means of statistical event study
Testing the investor sentiment proxy by OLS regression on VIX
4. Results
Identifying the Events
Testing maximum – minimum spread
Testing cross sectional dispersion
Testing kurtosis and skewness
Testing proxy of investor sentiment by means of statistical event study
Testing investor sentiment proxy by OLS regression on VIX
5. Discussion and Conclusions
Appendix
Bibliography
1. Introduction
This paper is primarily motivated by the question to what extent the sentiment of AEX investors is influenced by (negative) adjustments of credit rating agencies (hereafter CRAs) on the solvency of countries. In addressing this question, we present a study of investor sentiment. The study is based on investor predictions of the Amsterdam Stock Exchange (AEX index). In our research we investigate whether the predictions by the investors on the AEX index were affected by credit worthiness of sovereign credit rating news facts coming to the market by three major CRAs, namely Moodys, Standard&Poors and Fitch.Hence, this paper examines the importance and significance of CRAs as source of information for investors forecasting.
CRAs have considerable access to sovereign data and are therefore an important source of information for capital market participants (Duff and Einig, 2009). The influence of CRAs in the market becomes increasingly interesting. Because CRAs are concerned with the overall credit quality of sovereign bondissues and issuers, they focus not just on earnings performance.Moreover,credit ratings attempt to measure the probability of default; a continuous variable that changes as new information arrives according to Holthausen and Leftwich (1986).The question that arises is whether investors extract earnings related information from rating changes and adjust accordingly.
Moreover, in the present economic crisis it is interesting to investigate the roll of CRAs and hence the importance to investors on credit news that they bring to the market. The flow of news facts on credit solvency is crucial in markets where investor confidence is not particularly strong (e.g. crisis) (Kräussl, 2005). Investors’ behaviour is more volatile and hence influenced by market information. Even ifratingagenciesdo not behave pro-cyclically, their announcements may still trigger market jitters because many (institutional)investors (can) hold only investment-grade instruments (KaminskySchmukler, 2002).
Downgrading (or upgrading) sovereign debt below (or above) investment grade may have a drastic impact on prices because these rating changes can affect the pool of investments. The question is whether investors are aware of this phenomenon and hence adjust their expectations on the AEX.Our hypothesis is to test whetherinvestors’ sentiment is affected by news announcements by CRAs on sovereign credit ratings.
Both intuitive reasons and previous research help explain why the sentiment of investors is affected by the news announcements by CRAs. Primarily once we look at previous research, the role of CRAs can be interpreted as pro-cyclical behaviour. Rating agencies decide to downgrade (upgrade) a country when the prices of its financial instruments go down (up). Alternatively, the behaviour of prices in the days preceding rating and outlook changes could also reflect an anticipation effect. Market participants anticipate to the behaviour of rating and outlook changes, so markets discount those events. Hence, the investors may anticipate on these outlook changes announced by CRAs.Another study byStambaugh (2011), has investigated the relation between the government debt issues and whether this weighted negatively on the investor sentiment. In our study we focus most on the US and Europe that both depicted a period of extreme high volatility in the equity markets.
Multiple intuitive arguments assert a relationship between what CRAs aggravate and the reaction on investor sentiment. Firstly, upon a downgrade or negative outlook by the major CRAs on sovereign credit ratings, high volatility in the equity market can be observed; share prices fall, which in turn will have a negative impact on investor sentiment (Kaminskyet al, 2002). Moreover, lacking or negative growth outlooks by CRAs have negative effect on investor sentiment due to worsening market conditions such as consumer confidence. Another intuitive argument to support the relation between CRAs outlook changes and investor sentiment could be explained by a change in the liquidity in the market. In other words, once a CRA has downgraded the outlook of a country’s credit rating, interest payments will rise and hence the liquidity in the market will deteriorate. This will have a negative impact on the investor sentiment. These arguments are most important to investors in subsequent countries to which CRAs change the credit worthiness. However, the AEX is well intertwined globally, most importantly Europe and the United States. Evidence by Ferreira and Gama (2007) indicates that sovereign debt rating and credit outlook changes of one countryhave an asymmetric and economically significant effect on the stock market returns of other countriesover 1989–2003. Hence, CRA outlooks in foreign markets may have an effect on businesses traded on the AEX. Therefore, it is important to investigate whether the investors are aware of this risk and hence anticipate.
In answering the question to what extent the sentiment of AEX investors is influenced by (negative) adjustments of CRAs on credit outlook and rating adjustments of countries, we use a dataset that is provided to us by one of the major banks in the Netherlands. This dataset was constructed by a survey conducted under preferred investors. These investors all have a minimum of one hundred thousand euro invested in the market. The survey reveals the expected minimum and maximum closing price of the AEX in two weeks (sequence of the surveys) for each survey participant. Hence, the dataset providesus with a minimum – maximum spread on the expectations of investors. This spread can be used as a proxy on the investor sentiment. The minimum – maximum spread is expected to become larger in high volatile periods. Moreover, the distribution is expected to be non-normal, skewed to the minimum in comparison to the mean. Moreover, we look into other proxies of investor sentiment such as cross sectional dispersion within the point forecast and the skewness of the distribution around the CRA announcements.
In turn we make use of the credit solvency data of countries that is provided to us by the CRAs Moodys, Standard&Poors and Fitch. These CRAs all have their own grading scales based on down and upgrades just like their positive and negative outlook changes. We have categorized the down – and upgrade events for each CRA for the period December 2008 until February 2012.Moreover we have looked into the outlook changes by the CRAs for the same time horizon. We have a control group of 14 surveys (2 week periods) in which no announcements were made by the CRAs. In total there are 59 downgrades and 3 upgrades, 78 negative outlooks and 15 positive outlooks in our research horizon. These solvency changes are investigated for multiple countries that are directly or indirectly related to the AEX.
The theoreticalevidence and institutional background of CRAs aredescribed in section 2. The methodology concerning themaximum – minimum spread, cross-sectional dispersion, skewness and kurtosis as proxies for investor sentiment reacting on the rating changes of sovereign credit ratingsare reported in section 3. Moreover, the methodology for the event study based on statistical tests is described in this section. Section 4 describes the results of the relationship between the proxies and CRA announcements. The discussion on the research and results are described in section 5, which also contains the conclusion. An appendix describes the methodology tests in more detail.
2. Theoretical evidence and institutional background
We look into detail on the sovereign credit ratings of a specific sample of countries. CRAs assess sovereign credit ratings and assign a credit rating that is designed to measure the default risk only, not the risk ofinterest rate changes due, for example, to shifts in expectations of inflation. There are two important alternative views of how CRAs obtain information on the default risk of countries. One view is that the rating agencies have access to onlypublicly available information, and the agencies generally lag the market in processing that information. According to this view, if rating agencies base the rating changes on publicly available information, the efficient market hypothesis (EMH) predicts that stock prices will not adjust in response to the ratings change event (Brooks, 2004). Moreover, if this information is publically available,the CRA has the sole role to bundle the information and bring this through their existing channels into the market.
An alternative view is that CRAs are information specialists, which retrieve their information from multiple sources. These sources are not publically available and hence are very costly to retrieve. The selected CRAs have access to these channels and are hence the lowest cost providers of this information. Consequently, new information on the credit outlook of countries, assigned ratings, presented by the CRAs will have an effect on the attitude of investors that will anticipate to this (new) information. Hence, once there is an extent to which stock prices react to sovereign ratings implies either evidence against the semi-strong form EMH, or, the presence of some private information available only to ratings agencies that has, as a consequence, come into the public domain (Brooks, 2002).
Kaplan and Urwitzargue that the simple discriminant models based on publically available data too closely mimic the ratings of CRAs. This supports the claim that CRAs react on information that is already publically available. Moreover, anecdotal evidence according to Kaminskyet al,suggests that market participants do not try to anticipate to the actions of CRAs, but these agencies closely follow the market sentiment. Even though, according to the CRAs publications, the agencies claim that they have access to non-public information. Both Moody’s and Standard and Poor’s indicate that the rating review includes thorough investigations that are “probably” not publically accessible.
However, the results of Kaminskyet al are consistent with the findings in Reinhart (2001), who examines whether CRAs actions anticipated the crises of the 1990s. With a large sample of countries and crises, the author concludes that rating changes are far from being leading indicators of crises and have turned out to be lagging indicators of financial collapses. In contrast, the aftermath of rating changes is uneventful, with sovereign bond yield spreads and stock spreads mostly unchanged after the announcements and with both spreads maintaining the gains or losses observed in the days preceding the rating changes (Kaminsky et al, 2002).
Since the beginning of the crisis, CRAs have frequently been accused of timing their downgrades badly and of precipitating sudden negative shifts in investor consensus (Veron, 2011). However, according to Veron, it is infrequent that rating downgrades surprise markets – generally they follow degradations of market sentiment rather than precede it. According to previous research, once CRAs do anticipate, investors often do not give much attention. However, the investor sentiment has never been thoroughly investigated anticipating to such events.
Ever since the Euro-area has become volatile and the sovereign yields have come under pressure, Veron research suggests that ‘the biggest and most sudden shifts in investor sentiment have been triggered by new information from the policy sphere—such as, among others, the announcement by Greece of worse deficits than previously disclosed, the French-German Deauville declaration of October 18, 2010, or the Eurogroup President’s suggestion of a “re-profiling” on May 16, 2011’. In the research by Veron (2011) the author doubts whether CRA downgrades of Euro-area countries have had any impact on the sentiment of investors. The dataset that we are provided with, allows us to test these doubts. Clearly, we can agree that the policy signals have had demonstrable impact on risk perceptions of investors as was also proved by the BIS Quarterly review, 2010.
According to Moody’s in 2009, credit ratings are affected by the economic cycles. However, these credit ratings are, according to Moody’s, more stable and effective than market based indicators of credit worthiness. According to this CRA, replacing these credit ratings by market-based measures, the pro-cyclicality would be reinforced and hence worsening the financial market overall due to increased volatility.
Many opposing opinions and results become apparent once we look into the influence of CRAs on investor sentiment. There are clearly two sides to be taken, namely, either the CRAs have significant impact on the market and cause market jitters or CRAs are far behind the market and investors disregard the (new) information that the agencies provide. Nevertheless the two camps, this makes theresearch on the influence of CRAs on investor sentiment even more interesting and contemporary.
Multiple intuitive relations can be discussed between CRA activities on sovereign ratings and investor sentiment. Primarily, Reisen and Maltzan (1998) observe higher volatility in the equity markets after a credit rating event on a 30-day horizon. This increased volatility drives down share prices and triggers investor sentiment.
Brooks, Faff, Hillier and Hillier also report in 2002 that rating downgrades have a negative wealth impact on market returns. They found that a sovereign downgrade has a negative impact on both the domestic stock market and the dollar value of the country’s currency. Moreover, they have investigated the four major CRAs and concluded that only Standard&Poors and Fitch rating downgrades result in significant market falls. Finally, they proved that markets react more severely to multiple rating changes close after each other, interesting to our research since we have a dataset with a high density of rating announcements and downgrades.
Reisen et al investigatedin 1998 the market volatility after a negative announcement on sovereign ratings by the two leading CRAs (Moody’s and Standard&Poor). They found a significant raise in stock market volatility in the countries concerned with a negative announcement on their sovereign rating. They have attributed the significant impact of these sovereign ratings to investor herd behaviour and prudential regulation imposed on institutional investors (Reisen et al., 1998). These researchers indeed suggest that the CRAs described in their research have significant impact on the volatility in the market after downgrading the sovereign rating. They also suggest that the sovereign ratings system, as it is now, is a late warning system. Hence, the sovereign ratings system should be changed to an early warning system since the ratings, as it is now, contribute to a destabilizing international capital flow and reduce the benefit especially in the context of emerging markets according to Reisen et al.
Another relation between the CRA actions on sovereign credit ratings and investor sentiment is consumer confidence. The consumer confidence is affected once a CRA changes the sovereign credit outlook or rating. Hence, the negative growth outlooks and downgrades on sovereign credit ratings trigger consumer confidence. Moreover, during the financial crisis, the volatility in the western financial markets has contributed to a sharp decline in business and consumer confidence (Plosser, 2011). CRAs contribute to the panic among investors during crisis once downgrading, causing capital outflows and even higher spreads according to Reisen (2002). The sovereign ratings lag rather than lead and hence create additional consumer and investor unrest in the market.